Participants
This analytical cross-sectional study was performed on 51 patients with T2DM (17 males and 34 females) referring to Ophthalmology Center of Vali-e Asr Hospital, Zanjan, Iran. First, complete eye examinations was conducted on each case by a retina specialist and, accordingly, the patients with T2DM were classified into four groups: mild NPDR (mild non-proliferative diabetic retinopathy) (n = 12); severe NPDR (severe non-proliferative diabetic retinopathy) (n = 12); PDR (patients with proliferative diabetic retinopathy) (n = 12); and NRDP (non-retinopathy T2DM as the control group) (n=15). DR and its stage were diagnosed and certified by an eye specialist. Ethical approval was obtained from Research Committee of Zanjan University of Medical Sciences, Zanjan, Iran (ZUMS.REC.1395.259). Search related to the current study was conducted in accordance with the ethical principles stated in “Declaration of Helsinki”. Conventional dietary intake was assessed with the use of a 148-item semi-quantitative food frequency questionnaire (FFQ). The validity and reliability of the FFQ were described in detail another study (Mirmiran et al., 2010). The FFQ contains a list of Iranian food and drink items with standard portion sizes commonly consumed by Iranians. All the participants were asked to report their frequency of intake of a given serving of each item during the previous year. Then, consumption of vitamin A-containing food items was compared in different studied groups in order to control vitamin A intake as a confounding variable (28). The mean age of the participants was 60 years old. Inclusion criteria were as follows: having mean age of 60 (50–70) years old and having body mass index (BMI) <35 kg/m². The exclusion criteria were having nephropathy, heart disease or thyroid dysfunction, inflammatory diseases, cancers, and history of stroke, smoking and consuming dietary supplements.

Anthropometric and Biochemical Measurements
Anthropometric parameters such as height, weight, as well as waist and hip circumference were measured by WHO approved methods (29). For BMI evaluation, each participant, in light clothes and no shoes, was weighed (in kilogram, using a digital scale, to nearest 0.1 Kg) and measured for height (in meters, using a stadiometer, to the nearest 0.5 cm). BMI was calculated as weight in kg divided by height in m squared (kg/m²). Waist and hip circumferences were measured at the halfway between the lower rib to the iliac crest and bumps, respectively.
Blood samples of all the participants were taken after overnight fasting at 7–10 a.m. Serum was separated and frozen at -80°C until the analysis time. Fasting blood glucose, C-reactive protein, and lipid profiles (cholesterol, triglyceride, HDL, and LDL) were measured by enzymatic methods (Bionic Co. Kit; Tehran, Iran) and Hitachi biochemical autoanalyzer.

Serum RBP4 Levels
Serum RBP4 levels were determined by enzyme-linked immunosorbent assay (ELISA) using a commercial kit (Abcam, ab108897-Retinolbinding protein4 (RBP4) Human ELISA kit) with the sensitivity of 3.4 ng/mL, and the inter- and intra-assay variabilities were 3.1% and 5.4%, respectively.

Serum Vitamin A Levels
Serum vitamin A levels were determined by a C18 reverse-phase HPLC/uv isocratic system (Knauer, Germany) with 250 mm×4.6 mm column and 3μm particle size, using a procedure previously described (30) along with some modifications. All the used chemicals, methanol (cat. no. 1060072500), ethanol (cat. no. 1117272500), acetonitrile (cat. no. 1000302500), and distilled water (cat. no. 773218), were purchased from Merck. All-trans-retinol (cat. no. R7632) as external standard, retinyl acetate (cat. no. 46958) as internal standard, and hexane (cat. no. 110543) were obtained from Sigma-Aldrich. All the chemicals were HPLC purified grade. Stock solutions of retinol (20mg/L; 105 µmol/L) and retinyl acetate (20 mg/L; 70 µmol/L) were prepared in ethanol. All the stock solutions were stored at -20°C (maximum of three months). Work solutions of vitamin A (retinol) and internal standard were prepared on the daily basis. Briefly, serum was isolated from fasting blood samples by refrigerator centrifuge (3500g) for 10 min. Serum solution was allocated in a number of microtubes, covered with aluminum foil, and filled by N₂ gas to minimize light exposure and vitamin A oxidation. The internal standard (retinyl acetate) solution (100 µL), 150 µL of methanol, and 200 µL of ethanol were pipetted into a well-capped polypropylene tube; 200 µL of serum was added and the content was mixed on a vortex for 10 s. Then, 500 µL of hexane was added and the solution was vortexed for 60 s. Each sample solution was centrifuged (5000 RPM, 5 min, 4ºC) and supernatants were transferred to other microtubes. Repeatedly, the samples were centrifuged (5000 RPM) for 5 min and each isolated supernatant was added to the microtube of isolated supernatant of the previous step. The combined supernatant was evaporated to dryness under nitrogen gas at 45-50ºC. Then, the dried hexane phase was resuspended in 200 of methanol and vortexed for 10 s. Afterwards, 100 µL of the vortexed solution was injected into the HPLC machine, using a reverse-phase 3 µm column (250 mm [Length] and 4.6 mm [inner diameter]), at the isocratic flow rate of 0.8 mL/min; the pressure was 53 bar, and the column’s temperature was 30°C. The serum vitamin A (retinol) levels were detected by using an (all-trans-retinol)-retinyl acetate standard curve at 325 nm (30). To ensure reproducible retention on the column, the column was washed with methanol (1 mL/min for 60 min at 45°C) prior to loading the next sample.

Statistical Analysis
Statistical tests were performed by SPSS version 20 (SPSS Inc., Chicago Ill., USA). All the data were reported as mean ± SD. ANOVA and LSD post-hoc tests were run to compare differences between the groups. Kolmogorov-Smirnov test was performed to check for the normal distribution of the variables. Spearman and Pearson’s correlation coefficients were employed to investigate the relationship between different parameters. The levels of significance was set at P-value<0.05.

 

In total, 51 patients (17 men, 34 women) aged 48-73 years old were enrolled in the present study. Among those patients, 36(70%) were diagnosed with DR. Their basal characteristics as well as anthropometric and biochemical data after stratification into four groups are listed in Table 1. The mean age had no significant difference among all the groups. The levels of anthropometric indices (BMI and WHR) were similar in those groups. The plasma FBG levels were similar at high levels in all the groups, which indicated the groups were suffering from diabetes. As shown in Table 1, patients with severe NPDR had higher levels of TG and TC than the control group (P=0.059 and P=0.047, respectively). No significance difference was found in plasma HDL-c in four groups.
The data suggested that mean plasma RBP4 concentration was significantly higher in three different grades of DR than the control group (77.0±11.0, 81.7±10.9 and 88.3±11.9 vs. 71.4 ±12.3, respectively; P=0.004) (Table 1). For each unit (1 ng/mL) increase in the plasma level of RBP4, the adjusted risk of PDR would be increased by 18% (with the odd ratio of 1.184 (CI 95% (1.018 – 1.378) P=0.028).    
The plasma vitamin A levels were significantly deceased in DR subgroups compared with the control group suffering from diabetes. The ratio of plasma RBP4 to retinol in DR groups was found to be significantly higher than that in the control groups (0.21±0.06, 0.27±0.12 and 0.28±0.07 vs. 0.16±0.14, respectively; P=0.001) (Table1, Figure1). Post-hoc analysis indicated that, regarding RBP4, there was only a significant difference between control and PDR groups (P=0.003) and, in the case of vitamin A levels, there was significant differences between the control group and severe retinopathy and PDR groups (P=0.03 and P=0.016, respectively) (Table1).

Table 1. Baseline characteristics and biochemical parameters of DR groups and control (NRDP) patients

_{DR, Diabetic retinopathy; NRDP, Non-retinopathy diabetic patients; Data are expressed as mean; BMI, Body mass index; WHR, Wist-to-hip ratio;  FBS, Fast blood glucose; TG, Triglyceride; TC, Total cholesterol; LDL-C, Low density lipoprotein-cholesterol; HDL-C, High density lipoprotein- cholesterol; CRP, C-reactive protein; RBP4, Retinol binding protein4; SBP, Systolic blood pressure; DBP, Diastolic blood pressure; *, P<0.05 denotes statistically significant difference; † denotes post-hoc analysis which indicates significant difference of RBP4 between control and PDR groups (P=0.03) and significant differences of vitamin A between control, severe NPDR, and PDR groups (P= 0.03 and 0.016, respectively).}
 

Figure 1. Ratio of plasma RBP4-to-retinol versus control and three groups of diabetic retinopathy (DR). Control, Non-retinopathy diabetic patients; Mild, Mild non-proliferative diabetic retinopathy (NPDR); Severe, Severe non-proliferative diabetic retinopathy; PDR, Proliferative diabetic retinopathy. Plasma RBP4-to-retinol ratio in DR groups is significantly higher than that of the control group (*, P<0.05; **, P<0.01).

Discussion

Several studies have reported the link between RBP4 and diabetic retinopathy. However, most of these studies have not considered the status of vitamin A (retinol) levels, which is needed to maintain visual process, and the change of which may lead to ocular complications such as DR.
In the present study, the association of RBP4-to-vitamin A ratio with different grades or severity of DR was investigated in comparison with the patients with NRDM as the control group. The results indicated significant elevation in the plasma levels of TG and TC in the patients with different grades of DR compared with the control group, which was similar to that found in previous studies (3, 21). These investigations have shown that severity of DR is positively correlated with increasing triglyceride and plasma triglyceride level and so with plasma HDL-c level in T2DM patients. There was no correlation between the plasma RBP4 levels and biochemical and anthropometric parameters, which is similar to the results obtained by Akbay et al. (24).  
To the best of our knowledge, there is no study for assessing RBP4-to-vitamin A ratio in the various stages of DR. Our results indicated significant elevation of RBP4 level in serum of DR patient groups compared with NRDP control group. Similar observations have been made by some studies and they have suggested increased plasma RBP4 levels has a causative role in retinal degeneration and vision loss, which may lead to DR. Hence, they have demonstrated elevated RBP4 levels induce inflammation in human retinal endothelial cells through JNK and P38 MAPK signaling (10, 23, 31). However, plasma vitamin A levels were negatively related to severity of DR, the levels of which decreased in the patients with PDR and NPDR compared with NRDP (control group). Hence, our results represented RBP4-to-vitamin A ratio was significantly increased in correlation with severity of DR compared with the control group (Figure 1). It is clear that vitamin A is needed to maintain the visual process. Deceased level of vitamin A is associated with an adverse profile of oxidative stress, which has a role in the pathogenesis of DR, as suggested by some authors (32, 33). Therefore, at lower serum levels of vitamin A, the risk of DR and its severity could increase. 
In a case-control study on 287 patients suffering from diabetes, Li (22) described that elevated levels of RBP4 in plasma were associated with DR in Chinese patients with T2DM and suggested that lowering RBP4 could be a new strategy for treating T2DM with DR. This was supported by some other research in different populations (21, 23). Takebayashi et al. (23) found significantly elevated plasma RBP4 levels in the patients with PDR versus NPDR. In contrast, in another study (24), negative conclusion was made about the relationship of RBP4 and DR and it was explained that no effect of serum RBP4 was observed on DR or cardiovascular complications in T2DM. In an experimental study on transgenic mice, Du et al. (31) described that over-expressing RBP4 developed progressive retinal degeneration by a retinoid-independent mechanism. They suggested that elevation of RBP4 levels could be a risk factor for retinal damage and vision loss in patients with and without diabetes, regardless of vitamin A. Similarly, two other works (34, 35) have suggested retinol-free RBP4 (apo-RBP4) is a potent as retinol-bound RBP4 (holo-RBP4) in terms of inducing endothelial inflammation during microvascular complications of diabetes. However, Osman et al. (32) conducted a case-control study (on 25 healthy persons and 60 patients with DR) and concluded that the deficiency of vitamin A could affect the process of retinal cell proliferation.
Circulating RBP4 is an adipokine and is also secreted by the liver which transports vitamin A. RBP4 has been implicated in the pathogenesis of complications associated with insulin resistance such as DR.  what proportion of plasma RBP4 is secreted from liver, and what proportion from adipose tissue, is yet under discussion. Fedders et al. (36) used adeno-associated viruses (AAV) that comprise a liver-specific promoter to drive the expression of murine RBP4 in the livers of adult mice to test whether RBP4 impairs glucose homeostasis. Their findings indicated that liver-secreted RBP4 did not affect glucose homeostasis. However, these researchers suggested that plasma RBP4 was exclusively derived from hepatices. Nevertheless, several other studies (37, 38) have indicated that RBP4 which is derived from adipose tissue acts as an adipokine and contributes in metabolic syndrome mechanisms.
Totally, the main finding of the present study showed RBP4-to-vitamin A ratio increased in correlation with the severity of DR. This indicated increased plasma RBP4 levels and low level of plasma vitamin A in DR compared with NRDP. For this result, two hypotheses can be proposed; firstly, plasma RBP4 released from adipose tissue probably has high affinity binding and slower vitamin A delivery kinetics than RBP4 secreted from the liver. Secondly, plasma RBP4 released from the adipose tissue has lower affinity for STRA6 than the liver RBP4. STRA6 mediates the translocation of vitamin A from plasma RBP4 to the intracellular acceptor in peripheral tissues. Proving these hypotheses needs further verification by large-scale studies and more accurate assay methods which could determine the proportion of RBP4 secreted from the liver and adipose tissue, and RBP4 delivery kinetics in diabetic complications. The major limitation of this study was the sample size of the participants, which was small, and the strength was that it was the first time the plasma RBP4-to-vitamin A ratio was examined in relation to DR severity. No reports could be found to assess RBP-to-vitamin A ratio in the various stages of DR.
 
 

Conclusion

In conclusion, according to the data of the present study, RBP4-to-vitamin A ratio is positively correlated with DR severity, which indicated elevated plasma RBP4 levels and decreased plasma vitamin A levels in the process of DR. Since vitamin A is required to maintain visual process and decreased levels of vitamin A is associated with an adverse profile of oxidative stress, elevated RBP4 may have a role in the pathogenesis of DR complications. Therefore, regarding the importance of RBP4 delivery kinetic mechanisms in the process of DR, further experimental studies with larger scales and more accurate methodological issues are required.

Acknowledgements

The results described in the current study was extracted from an MSc thesis in the field of clinical biochemistry, under code ZUMS.REC.1395.259, which was approved by Zanjan Metabolic Diseases Research Center, Zanjan University of Medical Sciences, Iran, that financially supported the project. The authors thank all those who helped writing this paper, especially Department of Biochemistry, Zanjan University of Medical Sciences. We would also like to appreciate the efforts made by Mr.  Ehsan Nouri for providing fabulous technical assistance.

 

Conflicts of Interest

Authors declared no conflict of interests.